Currently, MIMO (multiple-input-multiple-output) is considered to be a key element of the air interface for high-speed wireless communications. MIMO can provide both diversity gain and multiplexing gain. MIMO enables simultaneous transmission of multiple streams, each stream being referred to as a layer. The number of transmit antennas, receive antennas, and layers are denoted by NT, NR, and R, respectively. The number of layers R is never greater than the number of transmit antennas NT, and is often smaller than or equal to the number of receive antennas NR.
Generally, MIMO assumes the use of a precoder, which is mathematically expressed as a left-multiplication of a layer signal vector (R×1) by a precoding matrix (NT×R). The precoding matrix is chosen from a pre-defined set of matrices, a so-called codebook as exemplified in Tables 1 and 2 for two and four transmit antennas, respectively. The r-th column vector of the precoding matrix represents the antenna spreading weight of the r-th layer.
TABLE 1Codebook for LTE UL (2-TX)Precoder MatrixNumber of layers νIndex120      1          2        ⁡      [                            1                                      1                      ]        1          2        ⁡      [                            1                          0                                      0                          1                      ]   1      1          2        ⁡      [                            1                                                  -            1                                ]  — 2      1          2        ⁡      [                            1                                      j                      ]  — 3      1          2        ⁡      [                            1                                                  -            j                                ]  — 4      1          2        ⁡      [                            1                                      0                      ]  — 5      1          2        ⁡      [                            0                                      1                      ]  —
TABLE 2Codebook for LTE UL (4-TX)PrecoderMatrixNumber of layers νIndex1234 0      1    2    ⁡      [                            1                                      1                                      1                                                  -            1                                ]        1    2    ⁡      [                            1                          0                                      1                          0                                      0                          1                                      0                                      -            j                                ]        1    2    ⁡      [                            1                          0                          0                                      1                          0                          0                                      0                          1                          0                                      0                          0                          1                      ]        1    2    ⁡      [                            1                          0                          0                          0                                      0                          1                          0                          0                                      0                          0                          1                          0                                      0                          0                          0                          1                      ]    1      1    2    ⁡      [                            1                                      1                                      j                                      j                      ]        1    2    ⁡      [                            1                          0                                      1                          0                                      0                          1                                      0                          j                      ]        1    2    ⁡      [                            1                          0                          0                                                  -            1                                    0                          0                                      0                          1                          0                                      0                          0                          1                      ]  —  2      1    2    ⁡      [                            1                                      1                                                  -            1                                                1                      ]        1    2    ⁡      [                            1                          0                                                  -            j                                    0                                      0                          1                                      0                          1                      ]        1    2    ⁡      [                            1                          0                          0                                      0                          1                          0                                      1                          0                          0                                      0                          0                          1                      ]  —  3      1    2    ⁡      [                            1                                      1                                                  -            j                                                            -            j                                ]        1    2    ⁡      [                            1                          0                                                  -            j                                    0                                      0                          1                                      0                                      -            1                                ]        1    2    ⁡      [                            1                          0                          0                                      0                          1                          0                                                  -            1                                    0                          0                                      0                          0                          1                      ]  —  4      1    2    ⁡      [                            1                                      j                                      1                                      j                      ]        1    2    ⁡      [                            1                          0                                                  -            1                                    0                                      0                          1                                      0                                      -            j                                ]        1    2    ⁡      [                            1                          0                          0                                      0                          1                          0                                      0                          0                          1                                      1                          0                          0                      ]  —  5      1    2    ⁡      [                            1                                      j                                      j                                      1                      ]        1    2    ⁡      [                            1                          0                                                  -            1                                    0                                      0                          1                                      0                          j                      ]        1    2    ⁡      [                            1                          0                          0                                      0                          1                          0                                      0                          0                          1                                                  -            1                                    0                          0                      ]  —  6      1    2    ⁡      [                            1                                      j                                                  -            1                                                            -            j                                ]        1    2    ⁡      [                            1                          0                                      j                          0                                      0                          1                                      0                          1                      ]        1    2    ⁡      [                            0                          1                          0                                      1                          0                          0                                      1                          0                          0                                      0                          0                          1                      ]  —  7      1    2    ⁡      [                            1                                      j                                                  -            j                                                            -            1                                ]        1    2    ⁡      [                            1                          0                                      j                          0                                      0                          1                                      0                                      -            1                                ]        1    2    ⁡      [                            0                          1                          0                                      1                          0                          0                                                  -            1                                    0                          0                                      0                          0                          1                      ]  —  8      1    2    ⁡      [                            1                                                  -            1                                                1                                      1                      ]        1    2    ⁡      [                            1                          0                                      0                          1                                      1                          0                                      0                          1                      ]        1    2    ⁡      [                            0                          1                          0                                      1                          0                          0                                      0                          0                          1                                      1                          0                          0                      ]  —  9      1    2    ⁡      [                            1                                                  -            1                                                j                                                  -            j                                ]        1    2    ⁡      [                            1                          0                                      0                          1                                      1                          0                                      0                                      -            1                                ]        1    2    ⁡      [                            0                          1                          0                                      1                          0                          0                                      0                          0                          1                                                  -            1                                    0                          0                      ]  — 10      1    2    ⁡      [                            1                                                  -            1                                                            -            1                                                            -            1                                ]        1    2    ⁡      [                            1                          0                                      0                          1                                                  -            1                                    0                                      0                          1                      ]        1    2    ⁡      [                            0                          1                          0                                      0                          0                          1                                      1                          0                          0                                      1                          0                          0                      ]  — 11      1    2    ⁡      [                            1                                                  -            1                                                            -            j                                                j                      ]        1    2    ⁡      [                            1                          0                                      0                          1                                                  -            1                                    0                                      0                                      -            1                                ]        1    2    ⁡      [                            0                          1                          0                                      0                          0                          1                                      1                          0                          0                                                  -            1                                    0                          0                      ]  — 12      1    2    ⁡      [                            1                                                  -            j                                                1                                                  -            j                                ]        1    2    ⁡      [                            1                          0                                      0                          1                                      0                          1                                      1                          0                      ]  —— 13      1    2    ⁡      [                            1                                                  -            j                                                j                                                  -            1                                ]        1    2    ⁡      [                            1                          0                                      0                          1                                      0                                      -            1                                                1                          0                      ]  —— 14      1    2    ⁡      [                            1                                                  -            j                                                            -            1                                                j                      ]        1    2    ⁡      [                            1                          0                                      0                          1                                      0                          1                                                  -            1                                    0                      ]  —— 15      1    2    ⁡      [                            1                                                  -            j                                                            -            j                                                1                      ]        1    2    ⁡      [                            1                          0                                      0                          1                                      0                                      -            1                                                            -            1                                    0                      ]  —— 16      1    2    ⁡      [                            1                                      0                                      1                                      0                      ]  ——— 17      1    2    ⁡      [                            1                                      0                                                  -            1                                                0                      ]  ——— 18      1    2    ⁡      [                            1                                      0                                      j                                      0                      ]  ——— 19      1    2    ⁡      [                            1                                      0                                                  -            j                                                0                      ]  ——— 20      1    2    ⁡      [                            0                                      1                                      0                                      1                      ]  ——— 21      1    2    ⁡      [                            0                                      1                                      0                                                  -            1                                ]  ——— 22      1    2    ⁡      [                            0                                      1                                      0                                      j                      ]  ——— 23      1    2    ⁡      [                            0                                      1                                      0                                                  -            j                                ]  ———
The precoding matrix usually consists of linearly-independent columns, and thus R is referred to as the rank of codebook. One purpose of a precoder is to match the precoding matrix with the channel so as to increase the received signal power and also to some extent reduce inter-layer interference, thereby improving the signal-to-interference-plus-noise-ratio (SINR) of each layer. Consequently, the precoder selection requires the transmitter to know the channel properties. Generally, the more accurate the channel information, the better the precoder matches.
In 3GPP LTE UL (3rd Generation Partnership Project's Long Term Evolution uplink), the precoder selection, which includes selection of both rank and precoding matrix, for use by the transmitter, e.g. UE (user equipment), is made at the receiver, e.g., eNodeB. Thus, it is not necessary for the receiver to feed channel information back to the transmitter.
Instead, it is necessary for the receiver to obtain the channel information so that a proper precoder selection can be made. This can be facilitated by the transmitter transmitting known signals such as DM-RS (demodulation reference signal) and SRS (sounding reference signal) in the case of LTE UL. An example is illustrated in FIG. 1 which illustrates an eNodeB 110, which serves a coverage area (or cell) 120 receiving uplink transmissions from UEs 130. The eNodeB 110 selects one precoder for use by UE 130-1, another precoder for UE 130-2 and yet another precoder for UE 130-3. Each UE 130 can facilitate the precoder selections by transmitting the known signal to the eNodeB 110. Both DM-RS and SRS are defined in the frequency domain and derived from Zadoff-Chu sequence.
It should be noted that DM-RS is precoded while SRS is not. Thus, the channel information obtained from DM-RS is the equivalent channel that the R layers experience, not the physical channel that the NT antennas experience. Mathematically, letting H denote the NR×NT physical channel matrix, W denote the NT×R preceding matrix, and E denote the NR×R equivalent channel, it follows thatE=HDW  (1)where D is the NT×NT diagonal matrix whose diagonal elements represent the inter-antenna gain/phase imbalance. Using the above notation, the equivalent channels for PUSCH (physical uplink shared channel), DM-RS and SRS denoted by EPUSCH, EDMRS and ESRS can be expressed asEPUSCH=HWEDMRS=HWHSRS=HD  (2)
Here it is assumed that there is no channel variation among the PUSCH, DM-RS and SRS and D is set to the identity matrix for PUSCH and DM-RS without loss of generality. It is also assumed that PUSCH and DM-RS experience the same channel. Also note that HSRS in (2) is directly obtained from SRS, and based on HSRS, the equivalent channel ESRS as a function of a hypothesized precoder, W can be obtained as ESRS=HSRSW.
Typically, the precoder is selected based on SRS, since it is more easily done with complete knowledge of channel, i.e., the physical channel, HD in (2). Based on the physical channel estimated based on SRS, the receiver chooses the best precoder and notifies the transmitter. One criteria for selecting the precoder is to maximize the throughput. For example, the effective SINR is calculated for each precoder, i.e., each selection of the rank and precoder matrix, the relevant throughput is calculated, and the precoder maximizing the throughput is selected. But it should be understood that precoder selection is subject to inter-antenna imbalance variation between measurement period and actual data transmission period.
Conventionally, the eNodeB measures the SRS transmitted from the UE. Based on the measurement, the eNodeB calculates the SINR of the SRS for each of the hypothesized precoder considered. A calculated SINR value corresponding to each of the hypothesized precoders is then directly used to select a MCS (modulation and coding scheme). Table 3 is a table that maps the MCSs to the SINRs. In this table, larger transport block sizes correspond to higher throughputs. Also, the MCSs are ordered such that higher MCSs correspond to higher throughputs.
TABLE 3TransportBlock Size6 RB25 RBRequired SINRMCSModulation(1.08 MHz)(5 MHz)for 10% BLER0QPSK152680−6.2681QPSK208904−5.1052QPSK2561096−4.3093QPSK3281416−3.2264QPSK4081800−2.1775QPSK5042216−1.3666QPSK6002600−0.6317QPSK71231120.4318QPSK80834961.1289QPSK93640081.98910QPSK103243922.7031116-QAM103243923.3421216-QAM119249683.9431316-QAM135257364.9961416-QAM154464565.8621516-QAM173672246.7041616-QAM180077367.2621716-QAM192879927.5291816-QAM215291448.8471916-QAM234499129.6322016-QAM26001068010.4532164-QAM26001068011.0002264-QAM27921144811.5952364-QAM29841257612.7132464-QAM32401353613.5892564-QAM34961411214.3202664-QAM36241526415.0892764-QAM37521584015.5542864-QAM43921883617.782
Note that selection of the MCS determines both the modulation (e.g., FSK, QPSK, QAM) and the throughput. For example, referring to FIG. 1, the SINR corresponding to a hypothesized precoder measured from the SRS transmitted by the UE 130-3 located at the edge of the cell 120 may be relatively low such as −3.5 dB and the SINR corresponding to the same hypothesized precoder measured from the UE 130-1 located closer to eNodeB 110 may be relatively high such as 12 dB. Using the conventional adaptation method, the eNodeB 110 would select MCS 2 for the UE 130-3 and MCS 22 for the UE 130-1. This means that the modulation used by the UEs 130-3 and 130-1 are QPSK and 64-QAM respectively, for the hypothesized precoder. Also if a 1.08 MHz bandwidth system is assumed, the corresponding throughputs would be 0.256 and 2.792 Mbps, respectively, for the hypothesized precoder. Such a MCS selection process is repeated for all other hypothesized precoders and at the end, the precoder that results in the highest throughput is chosen. The process of identifying a precoder that may result in the highest PUSCH throughput is called transmission mode adaptation. The precoder considered in such a process may of various ranks, and as such rank adaptation is also included as an element in transmission mode adaptation.
One problem of conventional transmission mode adaptation is that measurement (SRS for LTE UL) may experience a different power level compared to that of an actual data reception (PUSCH for LTE UL). This occurs because SRS and PUSCH may have different bandwidths and thus have different transmit power levels. As a result, the receiver may end up selecting a precoder that does not maximize the actual PUSCH throughput. This can be seen as the case where D in (2) has a positive real numbers larger or smaller than 1 as its diagonal elements. In the discussions below, examples of SRS having a higher receive power level than PUSCH (thus a higher gain level) will be used. In this case, a gain increase is likely to increase the effective SINR for each precoder and some of the precoders may reach the SINR for the highest MCS. Thus, the precoders that reach the highest MCS may achieve the maximum throughput.
In the absence of any gain increase, it does not matter which precoder is selected among those that reach the maximum throughput as any of these precoders also results in PUSCH to reach the maximum throughput. However, in the presence of gain increase, only one or some precoders may maximize the actual throughput. In the worst case, none of the precoders with the highest MCS may maximize the actual throughput.
Conventional transmission mode adaptation cannot select the precoder that maximizes the actual PUSCH throughput, since it always selects the precoder that maximizes the hypothetical throughput for the measurement. In the conventional transmission mode adaptation, the precoder is selected whose throughput is calculated first or last, depending on the implementation choice, e.g., its order of throughput calculation. The resulting performance tends to be similar to the performance of fixed precoder, i.e., performance without precoder selection, since it always selects a certain precoder, regardless of the actual throughput for PUSCH. This might also be the case, when the measurement experiences gain decrease and some of the precoders reach the lowest MCS and thus the same throughput.